Lightweight Combinational Machine Learning Algorithm for Sorting Canine Torso Radiographs

Tonima, M. A., Esfahani, F., Dehart, A., & Zhang, Y. Lightweight Combinational Machine Learning Algorithm for Sorting Canine Torso Radiographs. arXiv (2021). https://doi.org/10.48550/arXiv.2102.11385

This study addresses the lack of automation in veterinary radiograph sorting by developing a lightweight machine learning algorithm. Inspired by architectures like AlexNet, Inception, and SqueezeNet, the algorithm is designed to classify canine torso radiographs based on view and anatomy. The proposed model is more computationally efficient than SqueezeNet, while outperforming AlexNet, ResNet, DenseNet, and SqueezeNet in accuracy. This advancement offers potential for enhancing automation in veterinary diagnostic processes.